Network on Chip (NoC) architectures provide scalable platforms for designing Systems on Chip (SoC) with large number of cores. Developing products and applications using an NoC architecture offers many challenges and opportunities. A tool which can map an application or a set of applications to a given NoC architecture will be essential. In this thesis we first survey current techniques and we present our proposals for mapping and scheduling of concurrent applications to NoCs with multithreaded processors as computational resources. NoC platforms are basically a special class of Multiprocessor Embedded Systems (MPES). Conventional MPES architectures are mostly bus-based and, thus, are exposed to potential difficulties regarding scalability and reusability. There has been a lot of research on MPES development including work on mapping and scheduling of applications. Many of these results can also be applied to NoC platforms. Mapping and scheduling are known to be computationally hard problems. A large range of exact and approximate optimization algorithms have been proposed for solving these problems. The methods include Branch-and–Bound (BB), constructive and transformative heuristics such as List Scheduling (LS), Genetic Algorithms (GA) and various types of Mathematical Programming algorithms. Concurrent applications are able to capture a typical embedded system which is multifunctional. Concurrent applications can be executed on an NoC which provides a large computational power with multiple on-chip computational resources. Improving the time performances of concurrent applications which are running on Network on Chip (NoC) architectures is mainly correlated with the ability of mapping and scheduling methodologies to exploit the Thread Level Parallelism (TLP) of concurrent applications through the available NoC parallelism. Matching the architectural parallelism to the application concurrency for obtaining good performance-cost tradeoffs is  another aspect of the problem. Multithreading is a technique for hiding long latencies of memory accesses, through the overlapped execution of several threads. Recently, Multi-Threaded Processors (MTPs) have been designed providing the architectural infrastructure to concurrently execute multiple threads at hardware level which, usually, results in a very low context switching overhead. Simultaneous Multi-Threaded Processors (SMTPs) are superscalar processor architectures which adaptively exploit the coarse grain and the fine grain parallelism of applications, by simultaneously executing instructions from several thread contexts. In this thesis we make a case for using SMTPs and MTPs as NoC resources and show that such a multiprocessor architecture provides better time performances than an NoC with solely General-purpose Processors (GP). We have developed a methodology for task mapping and scheduling to an NoC with mixed SMTP, MTP and GP resources, which aims to maximize the time performance of concurrent applications and to satisfy their soft deadlines. The developed methodology was evaluated on many configurations of NoC-based platforms with SMTP, MTP and GP resources. The experimental results demonstrate that the use of SMTPs and MTPs in NoC platforms can significantly speed-up applications.

Des nombreuses applications participatives, commerciales et académiques se appuient sur des volontaires ("la foule") pour acquérir, désambiguiser et nettoyer des données. Ces applications participatives sont largement connues sous le nom de plates-formes de crowdsourcing où des amateurs peuvent participer à de véritables projets scientifiques ou commerciaux. Ainsi, des demandeurs sous-traitent des tâches en les proposant sur des plates-formes telles que Amazon MTurk ou Crowdflower. Puis, des participants en ligne sélectionnent et exécutent ces tâches, appelés microtasks, acceptant un micropaiement en retour. Ces plates-formes sont confrontées à des défis tels qu'assurer la qualité des réponses acquises, aider les participants à trouver des tâches pertinentes et intéressantes, tirer parti des compétences expertes parmi la foule, respecter les délais des tâches et promouvoir les participants qui accomplissent le plus de tâches. Cependant, la plupart des plates-formes ne modélisent pas explicitement les compétences des participants, ou se basent simplement sur une description en terme de mots-clés. Dans ce travail, nous proposons de formaliser les compétences des participants au moyen d'une structure hiérarchique, une taxonomie, qui permet naturellement de raisonner sur les compétences (détecter des compétences équivalentes, substituer des participants, ...). Nous montrons comment optimiser la sélection de tâches au moyen de cette taxonomie. Par de nombreuses expériences synthétiques et réelles, nous montrons qu'il existe une amélioration significative de la qualité lorsque l'on considère une structure hiérarchique de compétences au lieu de mots-clés purs. Dans une seconde partie, nous étudions le problème du choix des tâches par les participants. En effet, choisir parmi une interminable liste de tâches possibles peut s'avérer difficile et prend beaucoup de temps, et s’avère avoir une incidence sur la qualité des réponses. Nous proposons une méthode de réduction du nombre de propositions. L'état de l'art n'utilise ni une taxonomie ni des méthodes de classement. Nous proposons un nouveau modèle de classement qui tient compte de la diversité des compétences du participant et l'urgence de la tâche. À notre connaissance, nous sommes les premiers à combiner les échéances des tâches en une métrique d'urgence avec la proposition de tâches pour le crowdsourcing. Des expériences synthétiques et réelles montre que nous pouvons respecter les délais, obtenir des réponses de haute qualité, garder l'intérêt des participants tout en leur donnant un choix de tâches ciblé. / A large number of commercial and academic participative applications rely on a crowd to acquire, disambiguate and clean data. These participative applications are widely known as crowdsourcing platforms where amateur enthusiasts are involved in real scientific or commercial projects. Requesters are outsourcing tasks by posting them on online commercial crowdsourcing platforms such as Amazon MTurk or Crowdflower. There, online participants select and perform these tasks, called microtasks, accepting a micropayment in return. These platforms face challenges such as reassuring the quality of the acquired answers, assisting participants to find relevant and interesting tasks, leveraging expert skills among the crowd, meeting tasks' deadlines and satisfying participants that will happily perform more tasks. However, related work mainly focuses on modeling skills as keywords to improve quality, in this work we formalize skills with the use a hierarchical structure, a taxonomy, that can inherently provide with a natural way to substitute tasks with similar skills. It also takes advantage of the whole crowd workforce. With extensive synthetic and real datasets, we show that there is a significant improvement in quality when someone considers a hierarchical structure of skills instead of pure keywords. On the other hand, we extend our work to study the impact of a participant’s choice given a list of tasks. While our previous solution focused on improving an overall one-to-one matching for tasks and participants we examine how participants can choose from a ranked list of tasks. Selecting from an enormous list of tasks can be challenging and time consuming and has been proved to affect the quality of answers to crowdsourcing platforms. Existing related work concerning crowdsourcing does not use either a taxonomy or ranking methods, that exist in other similar domains, to assist participants. We propose a new model that takes advantage of the diversity of the parcipant's skills and proposes him a smart list of tasks, taking into account their deadlines as well. To the best of our knowledge, we are the first to combine the deadlines of tasks into an urgency metric with the task proposition for knowledge-intensive crowdsourcing. Our extensive synthetic and real experimentation show that we can meet deadlines, get high quality answers, keep the interest of participants while giving them a choice of well selected tasks.